screenpipe

screenpipe is a local-first memory layer: a desktop agent that records what you see and hear on your own machine, stores it locally, and makes it searchable by the AI tools you already use. Mediar, Inc. ships it for macOS, Windows and Linux, and describes the product as source-available rather than open source. Capture works through the operating system rather than through integrations. The app reads on-screen text primarily via accessibility APIs with OCR as a fallback, transcribes audio locally, and writes everything to a local store on your own disk. The vendor puts the cost of that at roughly 5 to 10 GB per month of continuous recording, configurable through retention settings. On macOS the app needs both screen recording and accessibility access granted in System Settings before it captures anything. The retrieval surface is where this row earns its integration score. screenpipe runs a local MCP server, installed with npx -y screenpipe-mcp, and exposes exactly two documented tools: search-content, which queries recorded screen text, audio transcriptions and input events with content-type, time-range and pagination parameters, and export-video. The vendor documents one-click installation into Claude Desktop from the app's own settings panel, plus manual stdio configuration for Claude Code, OpenAI Codex, Cursor and Warp, and generic instructions for other editors. An optional HTTP transport covers clients that cannot speak stdio. Under all of it sits a local REST API on port 3030, which is also what the MCP server talks to, so a pipe or a script can read the same history without going through an AI client at all. Pipes are the automation half: event-driven scripts installed from a pipe store or built with the SDK. Pricing is published, but not where you would look for it. The three tiers are listed on the About page: Basic at $25 per month for paid app access, Business at $50 per seat per month, and Enterprise at $150 per seat per month for managed deployments. No annual rate appears anywhere on the site, so the monthly figure is the whole ladder. The /pricing path redirects to an onboarding page carrying no prices at all, and the vendor's own FAQ still directs readers there. Currency is USD, stated in the site's own structured data on two separate pages. On privacy the architecture and the legal terms point in slightly different directions, and both are worth reading before adopting it. Data stays on the device by default, cloud model calls send only the query, and Ollama can be used for fully local inference. The terms state that customer content, output and personal information are not used to train models without separate written agreement. The same section then has the user consent, by using the service, to the creation and use of Deidentified Data and Analytics Data including to improve and train the vendor's own models. The public repository carries 21,150 stars and 2,122 forks, but its licence is not an OSI licence: personal, non-commercial and evaluation use can be free while commercial use requires a paid licence under the Screenpipe Commercial License.